diff --git a/.github/workflows/backend-docker-image-build.yml b/.github/workflows/backend-docker-image-build.yml
index f26e491..54e8976 100644
--- a/.github/workflows/backend-docker-image-build.yml
+++ b/.github/workflows/backend-docker-image-build.yml
@@ -4,7 +4,8 @@ on:
   push:
     branches:
       - "**"
-
+    paths:
+      - 'services/backend/**'
 jobs:
   backend_build:
     runs-on: ubuntu-latest
@@ -31,3 +32,14 @@ jobs:
           docker build -t $IMAGE .
           docker push $IMAGE
           cd ../..
+
+      - name: Deploy to Cloud Run
+        run: |
+          gcloud run deploy backend-service \
+            --source services/backend \
+            --region ${{ secrets.GCP_REGION }} \
+            --platform managed \
+            --allow-unauthenticated \
+            --memory 4Gi \
+            --timeout 3600s \
+            --set-env-vars "AIRFLOW_UID=5000,BASE_URL=https://www.khoury.northeastern.edu/,MAX_DEPTH=3,CONCURRENT_REQUESTS=10,DATA_FOLDER=scraped_data,MISTRAL_API_KEY=${{ secrets.MISTRAL_API_KEY }},MLFLOW_TRACKING_URI=${{ secrets.MLFLOW_TRACKING_URI }},BUCKET_NAME=${{ secrets.BUCKET_NAME }},RAW_DATA_FOLDER=raw_data,FAISS_INDEX_FOLDER=faiss_index,URLS_LIST= "https://www.khoury.northeastern.edu/""
diff --git a/.github/workflows/frontend-docker-image-build.yml b/.github/workflows/frontend-docker-image-build.yml
index 537cd3c..abd2475 100644
--- a/.github/workflows/frontend-docker-image-build.yml
+++ b/.github/workflows/frontend-docker-image-build.yml
@@ -4,6 +4,8 @@ on:
   push:
     branches:
       - "**"
+    paths:
+      - 'services/frontend/**'
 
 jobs:
   frontend_build:
@@ -31,3 +33,14 @@ jobs:
           docker build -t $IMAGE .
           docker push $IMAGE
           cd ../..
+
+      - name: Deploy to Cloud Run
+        run: |
+          gcloud run deploy frontend-service \
+            --source services/frontend \
+            --region ${{ secrets.GCP_REGION }} \
+            --platform managed \
+            --allow-unauthenticated \
+            --memory 1Gi \
+            --timeout 1800s \
+            --set-env-vars "API_URL=https://backend-service-273412-default.run.app/NuBot/"
diff --git a/.github/workflows/frontend1-docker-image-build.yml b/.github/workflows/frontend1-docker-image-build.yml
new file mode 100644
index 0000000..148f7ab
--- /dev/null
+++ b/.github/workflows/frontend1-docker-image-build.yml
@@ -0,0 +1,45 @@
+name: "build_reactfrontned_image"
+
+on:
+  push:
+    branches:
+      - "**"
+    paths:
+      - "services/frontend1/**"
+jobs:
+  backend_build:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: GCP Authentication
+        uses: google-github-actions/auth@v2
+        with:
+          credentials_json: "${{ secrets.GCP_KEY }}"
+
+      - name: Setup gcloud SDK
+        uses: google-github-actions/setup-gcloud@v2
+
+      - name: Docker login for Artifact Registry
+        run: |
+          gcloud auth configure-docker ${{ secrets.GCP_ARTIFACT_REGISTRY_REGION }}-docker.pkg.dev
+
+      - name: Build and Push Backend Image
+        run: |
+          cd services/frontend1
+          IMAGE=${{ secrets.GCP_ARTIFACT_REGISTRY_REGION }}-docker.pkg.dev/${{ secrets.GCP_PROJECT_ID }}/backend-nubot/react-service:latest
+          docker build -t $IMAGE .
+          docker push $IMAGE
+          cd ../..
+
+      - name: Deploy to Cloud Run
+        run: |
+          gcloud run deploy react-service \
+            --source services/frontend1 \
+            --region ${{ secrets.GCP_REGION }} \
+            --platform managed \
+            --allow-unauthenticated \
+            --memory 1Gi \
+            --timeout 3600s \
+            --set-env-vars "REACT_APP_API_URL=${{secrets.REACT_APP_API_URL}}"
diff --git a/.github/workflows/github-orchestration.yml b/.github/workflows/github-orchestration.yml
new file mode 100644
index 0000000..e69de29
diff --git a/.github/workflows/prefect_orchestraiton.yml b/.github/workflows/prefect_orchestraiton.yml
new file mode 100644
index 0000000..36116f7
--- /dev/null
+++ b/.github/workflows/prefect_orchestraiton.yml
@@ -0,0 +1,59 @@
+name: Deploy Prefect Flow to Cloud Run
+
+on:
+  push:
+    branches: ["**"] # Trigger on push to main (adjust as needed)
+
+jobs:
+  build-and-deploy:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v3
+
+      # Authenticate to Google Cloud using the service account JSON key
+      - name: GCP Authentication
+        uses: google-github-actions/auth@v2
+        with:
+          credentials_json: "${{ secrets.GCP_KEY }}"
+
+      - name: Setup gcloud SDK
+        uses: google-github-actions/setup-gcloud@v2
+
+      - name: Docker login for Artifact Registry
+        run: |
+          gcloud auth configure-docker ${{ secrets.GCP_ARTIFACT_REGISTRY_REGION }}-docker.pkg.dev
+
+      - name: Build Docker image
+        run: |
+          cd prefectWorkflows
+          IMAGE_URI=${{ secrets.GCP_ARTIFACT_REGISTRY_REGION }}-docker.pkg.dev/${{ secrets.GCP_PROJECT_ID }}/backend-nubot/prefect-scraper:latest
+          echo "Building image $IMAGE_URI"
+          docker build -t "$IMAGE_URI"   .
+          # Note: The context is the repository root (.), adjust path to Dockerfile if needed.
+
+      - name: Push Docker image to Artifact Registry
+        run: |
+          IMAGE_URI=${{ secrets.GCP_ARTIFACT_REGISTRY_REGION }}-docker.pkg.dev/${{ secrets.GCP_PROJECT_ID }}/backend-nubot/prefect-scraper:latest
+          docker push "$IMAGE_URI"
+        # After this step, the image is available in Artifact Registry for Cloud Run to use.
+
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.10"
+
+      - name: Install Prefect3
+        run: pip install --no-cache-dir "prefect>=3.2.4"
+
+      - name: Deploy via prefect deploy
+        run: |
+          cd prefectWorkflows
+          prefect deploy -n scraper-cron-deployment           # tell pool to use latest image
+
+      - name: Deploy Prefect flow
+        run: |
+          cd prefectWorkflows   # navigate to the folder containing prefect.yaml
+          prefect deploy -n scraperflow-deployment
+        # The -n flag ensures we deploy the specific deployment by name (optional if only one deployment in YAML).
+        # This command reads prefect.yaml and registers/updates the deployment in Prefect Cloud.
diff --git a/.gitignore b/.gitignore
index 71c9eff..166527d 100644
--- a/.gitignore
+++ b/.gitignore
@@ -177,6 +177,7 @@ mlflow-artifacts/
 # PyPI configuration file
 .pypirc
 
-*.html
+
 
 *.json
+!package.json
diff --git a/airflow/dags/dataflow/chunk_data.py b/airflow/dags/dataflow/chunk_data.py
index fc730d4..30f3ab4 100644
--- a/airflow/dags/dataflow/chunk_data.py
+++ b/airflow/dags/dataflow/chunk_data.py
@@ -11,7 +11,8 @@
 from dataflow.store_data import upload_faiss_index_to_bucket
 load_dotenv(override=True)
 BUCKET_NAME= os.getenv('BUCKET_NAME')
-GOOGLE_APPLICATION_CREDENTIALS=os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+from google.auth import default
+credentials, project = default()
 RAW_DATA_FOLDER= os.getenv('RAW_DATA_FOLDER')
 def chunk_data():
     # Load all JSON files from a directory
diff --git a/airflow/dags/dataflow/scraper.py b/airflow/dags/dataflow/scraper.py
index 9a0e5f9..d426ece 100644
--- a/airflow/dags/dataflow/scraper.py
+++ b/airflow/dags/dataflow/scraper.py
@@ -13,7 +13,8 @@
 BASE_URL = os.getenv('BASE_URL')
 MAX_DEPTH = int(os.getenv('MAX_DEPTH'))             # Maximum recursion depth (base URL is depth 0)
 CONCURRENT_REQUESTS = int(os.getenv('CONCURRENT_REQUESTS'))  # Maximum number of concurrent requests
-GOOGLE_APPLICATION_CREDENTIALS =os.getenv('GOOGLE_APPLICATION_CREDENTIALS ')
+from google.auth import default
+credentials, project = default()
 # Create folder for JSON data
 DATA_FOLDER = "scraped_data"
 if not os.path.exists(DATA_FOLDER):
diff --git a/airflow/dags/dataflow/store_data.py b/airflow/dags/dataflow/store_data.py
index d142be3..95ffa68 100644
--- a/airflow/dags/dataflow/store_data.py
+++ b/airflow/dags/dataflow/store_data.py
@@ -5,7 +5,8 @@
 BUCKET_NAME= os.getenv('BUCKET_NAME')
 RAW_DATA_FOLDER= os.getenv('RAW_DATA_FOLDER')
 FAISS_INDEX_FOLDER= os.getenv('FAISS_INDEX_FOLDER')
-GOOGLE_APPLICATION_CREDENTIALS=os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+from google.auth import default
+credentials, project = default()
 
 def get_blob_from_bucket():
     storage_client = Client()
diff --git a/prefectWorkflows/.dockerignore b/prefectWorkflows/.dockerignore
new file mode 100644
index 0000000..3168a00
--- /dev/null
+++ b/prefectWorkflows/.dockerignore
@@ -0,0 +1,3 @@
+*.env
+scraped_data/
+faiss_index/
\ No newline at end of file
diff --git a/prefectWorkflows/.env b/prefectWorkflows/.env
index 29ab14c..32d03ba 100644
--- a/prefectWorkflows/.env
+++ b/prefectWorkflows/.env
@@ -6,4 +6,6 @@ DATA_FOLDER = "scraped_data"
 BUCKET_NAME=scraped_raw_data_nubot
 RAW_DATA_FOLDER=raw_data
 FAISS_INDEX_FOLDER=faiss_index
-GOOGLE_APPLICATION_CREDENTIALS="E:/gcpkeys/nubot/nubot-nikhil-6adeee091d55.json"
\ No newline at end of file
+GOOGLE_APPLICATION_CREDENTIALS="E:/gcpkeys/nubot/nubot-nikhil-6adeee091d55.json"
+PREFECT_API_KEY=pnu_mRGcrBkC9qyFbwGfgrVbjbOoL7WIZ411TKYp
+PREFECT_API_URL="https://api.prefect.cloud/api/accounts/806f2e07-5063-4fbe-9b46-0545ad5de2d1/workspaces/acdf9e9e-8a55-446a-ac46-80a3f843d8b6"
diff --git a/prefectWorkflows/Dockerfile b/prefectWorkflows/Dockerfile
new file mode 100644
index 0000000..dfc7b42
--- /dev/null
+++ b/prefectWorkflows/Dockerfile
@@ -0,0 +1,25 @@
+# Start from a lightweight Python image (use the appropriate Python version)
+FROM python:3.10-slim
+
+# Set working directory in container
+WORKDIR /app
+
+# Install Python dependencies.
+# If you have a requirements.txt, copy and install it:
+COPY requirements.txt . 
+
+RUN pip install --no-cache-dir -r requirements.txt
+
+# (Alternatively, directly install Prefect and any needed libraries)
+# RUN pip install prefect==3.1.10
+
+# Copy the Prefect flow code and the dataflow module into the image
+COPY . .
+
+
+# Ensure Python can find the 'dataflow' module (add /app to PYTHONPATH)
+ENV PYTHONPATH="/app:${PYTHONPATH}"
+
+# (Optional) Set a default command (Prefect Cloud will override this when submitting the flow run)
+# By default, do nothing or use a generic command. Prefect Cloud's work pool will specify the entrypoint at runtime.
+CMD ["python", "-c", "print('Container built for Prefect flow execution')"]
diff --git a/prefectWorkflows/dataflow/chunk_data.py b/prefectWorkflows/dataflow/chunk_data.py
index 643dcba..fc35e37 100644
--- a/prefectWorkflows/dataflow/chunk_data.py
+++ b/prefectWorkflows/dataflow/chunk_data.py
@@ -11,7 +11,20 @@
 from dataflow.store_data import upload_faiss_index_to_bucket
 load_dotenv(override=True)
 BUCKET_NAME= os.getenv('BUCKET_NAME')
-GOOGLE_APPLICATION_CREDENTIALS=os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+from google.auth import default
+from google.oauth2 import service_account
+
+# Try to get credentials - works in both Docker and Cloud Run
+try:
+    # First try Application Default Credentials (works in Cloud Run)
+    credentials, project = default()
+except Exception:
+    # Fall back to explicit credentials file (for Docker)
+    credentials_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
+    if credentials_path:
+        credentials = service_account.Credentials.from_service_account_file(credentials_path)
+    else:
+        raise Exception("No credentials available")
 RAW_DATA_FOLDER= os.getenv('RAW_DATA_FOLDER')
 def chunk_data():
     # Load all JSON files from a directory
@@ -52,4 +65,4 @@ def chunk_data():
 
 if __name__=="__main__":
     chunk_data()
-    
+    upload_faiss_index_to_bucket()
\ No newline at end of file
diff --git a/prefectWorkflows/dataflow/rag_model.py b/prefectWorkflows/dataflow/rag_model.py
deleted file mode 100644
index ffcfa6c..0000000
--- a/prefectWorkflows/dataflow/rag_model.py
+++ /dev/null
@@ -1,149 +0,0 @@
-from functools import lru_cache
-from langchain import hub
-from langchain_core.documents import Document
-from langgraph.graph import START, StateGraph
-from typing_extensions import List, TypedDict
-from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain.chat_models import init_chat_model
-from langchain_community.vectorstores import FAISS
-import getpass
-import os
-from dotenv import load_dotenv
-import mlflow
-import time
-from langfair.auto import AutoEval
-import asyncio
-# Load the FAISS index
-from google.cloud.storage import Client
-import tempfile
-import os
-load_dotenv(override=True)
-mlflow.langchain.autolog()
-MLFLOW_TRACKING_URI =os.environ.get("MLFLOW_TRACKING_URI")
-MISTRAL_API_KEY = os.getenv("MISTRAL_API_KEY")
-FAISS_INDEX_FOLDER= os.getenv('FAISS_INDEX_FOLDER')
-mlflow.set_tracking_uri(MLFLOW_TRACKING_URI)  # Remote MLflow Server
-mlflow.set_experiment("rag_experiment")
-if not os.environ.get("MISTRAL_API_KEY"):
-  os.environ["MISTRAL_API_KEY"] = getpass.getpass("Enter API key for Mistral AI: ")
-
-@lru_cache(maxsize=None)
-def get_llm():
-    llm = init_chat_model("mistral-large-latest", model_provider="mistralai")
-    return llm
-
-@lru_cache(maxsize=None)
-def get_prompt():
-# Define prompt for question-answering
-    prompt = hub.pull("rlm/rag-prompt")
-    return prompt
-
-
-# Define state for application
-class State(TypedDict):
-    question: str
-    context: List[Document]
-    answer: str
-
-
-@lru_cache(maxsize=None)
-def load_embeddings():
-    embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
-    return embeddings
-
-
-
-# Initialize GCS client
-storage_client = Client()
-bucket=storage_client.bucket(os.getenv('BUCKET_NAME'))
-embeddings=load_embeddings()
-if not os.path.exists(FAISS_INDEX_FOLDER):
-    os.makedirs(FAISS_INDEX_FOLDER, exist_ok=True)
-# Create a temporary directory
-# Download FAISS index files from bucket to FAISS_INDEX_FOLDER directory
-for blob in bucket.list_blobs(prefix=FAISS_INDEX_FOLDER):
-    # Extract just the filename from the full path
-    filename = os.path.basename(blob.name)
-    local_path = os.path.join(FAISS_INDEX_FOLDER, filename)
-    blob.download_to_filename(local_path)
-
-# Load FAISS index from directory
-vector_store = FAISS.load_local(FAISS_INDEX_FOLDER, embeddings, allow_dangerous_deserialization=True)
-# Define application steps
-def retrieve(state: State):
-    with mlflow.start_run(nested=True, run_name="retrieval"):
-        start_time = time.time()
-        retrieved_docs = vector_store.similarity_search(state["question"])
-        retrieval_time = time.time() - start_time
-    
-        # Extract only metadata
-        doc_metadata = [{"doc_id": doc.metadata.get("id", i), "source": doc.metadata.get("source", "unknown")}
-                        for i, doc in enumerate(retrieved_docs)]
-        
-        # Log metadata instead of full documents
-        mlflow.log_metric("retrieval_time", retrieval_time)
-        mlflow.log_param("retrieved_docs_count", len(retrieved_docs))
-        mlflow.log_dict(doc_metadata, "retrieved_docs.json")
-
-    return {"context": retrieved_docs}
-
-# Initialize LLM once and store in a global variable
-llm = get_llm()
-# Initialize prompt once and store in a global variable
-prompt = get_prompt()
-def generate(state: State):
-    with mlflow.start_run(nested=True, run_name="generation"):
-        start_time = time.time()
-        docs_content = "\n\n".join(doc.page_content for doc in state["context"])
-        token_count = len(docs_content.split()) 
-        # Use the global prompt instance
-        mlflow.log_param("retrieved_tokens", token_count)
-        mlflow.log_param("context_length", len(docs_content))
-        messages = prompt.invoke({"question": state["question"], "context": docs_content})
-        response = llm.invoke(messages)
-        generation_time = time.time() - start_time
-        
-        # Log LLM generation performance
-        mlflow.log_metric("generation_time", generation_time)
-        mlflow.log_param("response_length", len(response.content.split()))
-        mlflow.log_param("model_name", "mistral-large-latest")
-
-        # Save response
-        # with open("response.txt", "w") as f:
-        #     f.write(response.content)
-        # mlflow.log_artifact("response.txt")
-
-    return {"answer": response.content}
-
-
-def generateResponse(query):
-# Compile application and test
-    try:
-         with mlflow.start_run(run_name="RAG_Pipeline"):
-            mlflow.log_param("query", query)
-            graph_builder = StateGraph(State).add_sequence([retrieve, generate])
-            graph_builder.add_edge(START, "retrieve")
-            graph = graph_builder.compile()
-            response = graph.invoke({"question": f"{query}"})
-            mlflow.log_param("final_answer", response["answer"])
-            return response["answer"]
-    except Exception as e:
-        mlflow.log_param("error", str(e))
-        raise Exception(e)
-    
-async def checkModel_fairness():
-    auto_object = AutoEval(
-        prompts=["tell me about khoury"], 
-        langchain_llm=llm,
-        # toxicity_device=device # uncomment if GPU is available
-    )
-    results = await auto_object.evaluate()
-    print(results['metrics'])
-    
-if __name__ == "__main__":
-
-    query=input("generate query")
-    response=generateResponse(query)
-    print(response)
-    #uncomment and enter prompts for model fairness and there is a limitation on api key
-    # asyncio.run(checkModel_fairness())
\ No newline at end of file
diff --git a/prefectWorkflows/dataflow/scraper.py b/prefectWorkflows/dataflow/scraper.py
index 9a0e5f9..4676a4a 100644
--- a/prefectWorkflows/dataflow/scraper.py
+++ b/prefectWorkflows/dataflow/scraper.py
@@ -6,14 +6,30 @@
 import re
 from urllib.parse import urljoin, urlparse
 from dotenv import load_dotenv
-
+import hashlib
 from dataflow.store_data import upload_many_blobs_with_transfer_manager
 load_dotenv(override=True)
 # Configuration
-BASE_URL = os.getenv('BASE_URL')
+URLS_LIST=list(os.getenv('URLS_LIST','').split(","))
+
+# BASE_URL ="" #URLS_LIST[0]#os.getenv('BASE_URL')
 MAX_DEPTH = int(os.getenv('MAX_DEPTH'))             # Maximum recursion depth (base URL is depth 0)
 CONCURRENT_REQUESTS = int(os.getenv('CONCURRENT_REQUESTS'))  # Maximum number of concurrent requests
-GOOGLE_APPLICATION_CREDENTIALS =os.getenv('GOOGLE_APPLICATION_CREDENTIALS ')
+
+from google.auth import default
+from google.oauth2 import service_account
+
+# Try to get credentials - works in both Docker and Cloud Run
+try:
+    # First try Application Default Credentials (works in Cloud Run)
+    credentials, project = default()
+except Exception:
+    # Fall back to explicit credentials file (for Docker)
+    credentials_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
+    if credentials_path:
+        credentials = service_account.Credentials.from_service_account_file(credentials_path)
+    else:
+        raise Exception("No credentials available")
 # Create folder for JSON data
 DATA_FOLDER = "scraped_data"
 if not os.path.exists(DATA_FOLDER):
@@ -24,7 +40,8 @@ def safe_filename(url):
     parsed = urlparse(url)
     path = parsed.path.strip('/') or 'index'
     filename = re.sub(r'[^A-Za-z0-9_\-]', '_', path) + ".json"
-    return os.path.join(DATA_FOLDER, filename)
+    url_hash = hashlib.md5(url.encode()).hexdigest()[:8]
+    return os.path.join(DATA_FOLDER, f"{filename}_{url_hash}.json")
 
 async def fetch(session, url, semaphore):
     """Fetch the content of the URL asynchronously."""
@@ -39,7 +56,7 @@ async def fetch(session, url, semaphore):
         print(f"Error fetching {url}: {e}")
         return None
 
-async def async_scrape(url, depth=0, session=None, semaphore=None):
+async def async_scrape(url,BASE_URL, depth=0, session=None, semaphore=None):
     """Recursively scrape pages asynchronously and store in JSON format."""
     if depth > MAX_DEPTH:
         return
@@ -81,26 +98,33 @@ async def async_scrape(url, depth=0, session=None, semaphore=None):
         next_url = urljoin(url, link['href'])
         if urlparse(next_url).netloc == urlparse(BASE_URL).netloc:
             next_url = next_url.split('#')[0]  # Remove fragments
-            tasks.append(async_scrape(next_url, depth + 1, session, semaphore))
+            tasks.append(async_scrape(next_url,BASE_URL, depth + 1, session, semaphore))
 
     if tasks:
         await asyncio.gather(*tasks)
     
 
-async def scrape_and_load():
+async def scrape_and_load(CURRENT_URl):
     """Main function to initiate scraping."""
     semaphore = asyncio.Semaphore(CONCURRENT_REQUESTS)
     
     async with aiohttp.ClientSession() as session:
-        await async_scrape(BASE_URL, depth=0, session=session, semaphore=semaphore)
+        await async_scrape(CURRENT_URl,BASE_URL=CURRENT_URl, depth=0, session=session, semaphore=semaphore)
     
 
 def scrape_and_load_task():
-    asyncio.run(scrape_and_load())
+    for url in URLS_LIST:
+        BASE_URL=url
+        asyncio.run(scrape_and_load(BASE_URL))
+        print("*"*15)
+        print(f"scraping {url} done")
+        print("*"*15)
+
     upload_many_blobs_with_transfer_manager()
     return
 
 
 if __name__ == '__main__':
-    asyncio.run(scrape_and_load())
-    upload_many_blobs_with_transfer_manager()
\ No newline at end of file
+    scrape_and_load_task()
+    # asyncio.run(scrape_and_load())
+    # upload_many_blobs_with_transfer_manager()
\ No newline at end of file
diff --git a/prefectWorkflows/dataflow/store_data.py b/prefectWorkflows/dataflow/store_data.py
index 039dc22..e89ad85 100644
--- a/prefectWorkflows/dataflow/store_data.py
+++ b/prefectWorkflows/dataflow/store_data.py
@@ -5,8 +5,20 @@
 BUCKET_NAME= os.getenv('BUCKET_NAME')
 RAW_DATA_FOLDER= os.getenv('RAW_DATA_FOLDER')
 FAISS_INDEX_FOLDER= os.getenv('FAISS_INDEX_FOLDER')
-GOOGLE_APPLICATION_CREDENTIALS=os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+from google.auth import default
+from google.oauth2 import service_account
 
+# Try to get credentials - works in both Docker and Cloud Run
+try:
+    # First try Application Default Credentials (works in Cloud Run)
+    credentials, project = default()
+except Exception:
+    # Fall back to explicit credentials file (for Docker)
+    credentials_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
+    if credentials_path:
+        credentials = service_account.Credentials.from_service_account_file(credentials_path)
+    else:
+        raise Exception("No credentials available")
 def get_blob_from_bucket():
     storage_client = Client()
     bucket = storage_client.bucket(BUCKET_NAME)
@@ -19,7 +31,7 @@ def get_blob_from_bucket():
 
 
 def upload_many_blobs_with_transfer_manager(
-    
+    workers=8
 ):
     """Upload every file in a list to a bucket, concurrently in a process pool.
 
diff --git a/prefectWorkflows/prefect.yaml b/prefectWorkflows/prefect.yaml
index ab669e5..879a8bd 100644
--- a/prefectWorkflows/prefect.yaml
+++ b/prefectWorkflows/prefect.yaml
@@ -1,47 +1,20 @@
-# Welcome to your prefect.yaml file! You can use this file for storing and managing
-# configuration for deploying your flows. We recommend committing this file to source
-# control along with your flow code.
+# Prefect deployment configuration for the scraper_flow
+name: scraper-flow-project # Name of the project (can be any identifier for your reference)
+prefect-version: 3.1.10 # Prefect version to use for this deployment (match your Prefect 3.x version)
 
-# Generic metadata about this project
-name: prefectWorkflows
-prefect-version: 3.2.15
-
-# build section allows you to manage and build docker images
-build:
-
-# push section allows you to manage if and how this project is uploaded to remote locations
-push:
-
-# # pull section allows you to provide instructions for cloning this project in remote locations
-# pull:
-# - prefect.deployments.steps.git_clone:
-#     repository: https://github.com/Nikhil-Kudupudi/NUBot.git
-#     branch: gcs_bucket
-#     access_token:
-
-# the deployments section allows you to provide configuration for deploying flows
 deployments:
-- name:
-  version:
-  tags: []
-  description:
-  schedule: {}
-  flow_name:
-  entrypoint:
-  parameters: {}
-  work_pool:
-    name:
-    work_queue_name:
-    job_variables: {}
-- name: default
-  version:
-  tags: []
-  concurrency_limit:
-  description:
-  entrypoint: scraper_flow.py:scraperflow
-  parameters: {}
-  work_pool:
-    name: nubot_dataflow
-    work_queue_name:
-    job_variables: {}
-  schedules: []
+  - name: scraperflow-deployment # Name of this deployment (appears in Prefect UI)
+    description: "Scrapes all URLs and segments data every Saturday at 9:00 UTC"
+    entrypoint: scraper_flow.py:scraper_flow # Entry point to the flow: "<script path>:<flow function>"
+    # Cron schedule for every Saturday at 9:00 AM
+    schedule:
+      cron: "0 9 * * 6" # Cron expression for Saturday 09:00 (UTC)&#8203;:contentReference[oaicite:3]{index=3}
+      timezone: "UTC" # Timezone for the schedule (adjust if needed)
+    parameters: {} # Default parameters (empty since this flow has none)
+    work_pool:
+      name: "my-cloud-run-pool" # Name of the push work pool for Cloud Run
+      work_queue_name: "default" # Work queue (use "default" or as configured in the pool)
+    tags: [] # (Optional) any tags for the deployment
+    # (Optional) infrastructure overrides can be specified if needed:
+    # infra_overrides:
+    #   image: "us-east1-docker.pkg.dev/<YOUR_GCP_PROJECT>/<YOUR_AR_REPOSITORY>/prefect-scraper:latest"
diff --git a/prefectWorkflows/requirements.txt b/prefectWorkflows/requirements.txt
new file mode 100644
index 0000000..c4a2008
--- /dev/null
+++ b/prefectWorkflows/requirements.txt
@@ -0,0 +1,17 @@
+python-dotenv
+requests
+transformers==4.48.0
+sentence-transformers
+torch
+faiss-cpu
+mlflow
+langchain[mistralai]
+langchain-community
+langgraph
+google-cloud-storage
+datasets
+gcsfs
+beautifulsoup4
+aiohttp
+langchain-huggingface
+langfair
diff --git a/prefectWorkflows/scraper_flow.py b/prefectWorkflows/scraper_flow.py
index b82f9a8..7521eee 100644
--- a/prefectWorkflows/scraper_flow.py
+++ b/prefectWorkflows/scraper_flow.py
@@ -1,32 +1,36 @@
-from prefect import flow, task
-from dataflow.scraper import  scrape_and_load_task 
-from dataflow.chunk_data import chunk_data
-
+from prefect import flow, task, get_run_logger
+# Import the supporting functions from dataflow module
+from dataflow.scraping import scrape_and_load_task  # adjust import to actual module path
+from dataflow.processing import chunk_data
 
+# Define Prefect tasks
 @task
 def scrape_all_urls_task():
-    # If scrape_all_urls is an imported function, call it here and return the result
-    return scrape_and_load_task()  # or return the relevant data
+    """Task to scrape all URLs and load raw data."""
+    logger = get_run_logger()
+    logger.info("Starting scrape_all_urls_task...")
+    data = scrape_and_load_task()  # call the helper function to scrape and load data
+    logger.info(f"Scraped data: {len(data)} items.")
+    return data
+
 @task
-def dataSegmentation():
-    return chunk_data()
+def dataSegmentation(data):
+    """Task to segment the scraped data into chunks."""
+    logger = get_run_logger()
+    logger.info("Starting dataSegmentation task...")
+    segments = chunk_data(data)  # call helper to chunk the data
+    logger.info(f"Segmented data into {len(segments)} chunks.")
+    return segments
 
-@flow(log_prints=True)
-def scraperflow():
-    # Use the tasks within the flow
-    scrape_all_urls_task()
-    dataSegmentation()
+@flow
+def scraper_flow():
+    """Prefect flow to orchestrate scraping and data segmentation."""
+    # Run the scraping task and then pass its result into the segmentation task
+    raw_data = scrape_all_urls_task()
+    segmented = dataSegmentation(raw_data)
+    # (Optional) do something with segmented data, e.g., save or return
+    return "done"
 
 if __name__ == "__main__":
-# # Run the flow
-## for cloud
-    # scraperflow.deploy(name="my-first-deployment",
-    #                    work_pool_name="dataflow",
-    #                    image='prefecthq/prefect:2-python3.10',
-    #                    push=True
-    #                   )
-    try:
-        scraperflow()
-    except Exception as e:
-        print(e)
-
+    # This allows testing the flow locally by running this script
+    scraper_flow()
diff --git a/requirements.txt b/run-requirements.txt
similarity index 100%
rename from requirements.txt
rename to run-requirements.txt
diff --git a/services/backend/.dockerignore b/services/backend/.dockerignore
index 58329ad..c618940 100644
--- a/services/backend/.dockerignore
+++ b/services/backend/.dockerignore
@@ -1,3 +1,7 @@
 *.env
 __pycache__/
-scraped_data/
\ No newline at end of file
+scraped_data/
+mlruns/
+mlflow.db/
+mlartifacts/
+logs/
\ No newline at end of file
diff --git a/services/backend/.env b/services/backend/.env
index 7cc284f..60515d6 100644
--- a/services/backend/.env
+++ b/services/backend/.env
@@ -5,6 +5,7 @@ CONCURRENT_REQUESTS = 10
 DATA_FOLDER = "scraped_data"
 MISTRAL_API_KEY="1UTJndGP95gDBZopQkDojiZ5bCzSJG9p"
 MLFLOW_TRACKING_URI="http://localhost:5000"
+URLS_LIST= "https://www.khoury.northeastern.edu/"
 PORT=8080
 HOST=127.0.0.1
 BUCKET_NAME=scraped_raw_data_nubot
diff --git a/services/backend/Dockerfile b/services/backend/Dockerfile
index b6a91f8..a91ae59 100644
--- a/services/backend/Dockerfile
+++ b/services/backend/Dockerfile
@@ -16,7 +16,8 @@ WORKDIR /app
 
 # Pre-copy requirements separately for Docker cache efficiency
 COPY runtime-requirements.txt .
-
+RUN pip install --upgrade pip
+RUN pip install "huggingface_hub[hf_xet]"
 # Install Python dependencies
 RUN pip install --no-cache-dir -r runtime-requirements.txt
 
diff --git a/services/backend/main.py b/services/backend/main.py
index 2a45b98..d2f578d 100644
--- a/services/backend/main.py
+++ b/services/backend/main.py
@@ -51,6 +51,6 @@ def post(self):
            
 
 if __name__=="__main__":
-    PORT=os.getenv('PORT')
-    HOST=os.getenv('HOST')
-    app.run(host=HOST,port=PORT,debug=True)
\ No newline at end of file
+    PORT=os.getenv('PORT', 8080)
+
+    app.run(host='0.0.0.0',port=PORT,debug=True)
\ No newline at end of file
diff --git a/services/backend/runtime-requirements.txt b/services/backend/runtime-requirements.txt
index 4ff0497..953f6d0 100644
--- a/services/backend/runtime-requirements.txt
+++ b/services/backend/runtime-requirements.txt
@@ -13,6 +13,7 @@ langchain-community
 langgraph
 google-cloud-storage
 datasets
+gcsfs
 beautifulsoup4
 aiohttp
 flask-cors
diff --git a/services/backend/src/dataflow/chunk_data.py b/services/backend/src/dataflow/chunk_data.py
index 7a43414..265ce1a 100644
--- a/services/backend/src/dataflow/chunk_data.py
+++ b/services/backend/src/dataflow/chunk_data.py
@@ -11,7 +11,20 @@
 from store_data import upload_faiss_index_to_bucket
 load_dotenv(override=True)
 BUCKET_NAME= os.getenv('BUCKET_NAME')
-GOOGLE_APPLICATION_CREDENTIALS=os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+from google.auth import default
+from google.oauth2 import service_account
+
+# Try to get credentials - works in both Docker and Cloud Run
+try:
+    # First try Application Default Credentials (works in Cloud Run)
+    credentials, project = default()
+except Exception:
+    # Fall back to explicit credentials file (for Docker)
+    credentials_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
+    if credentials_path:
+        credentials = service_account.Credentials.from_service_account_file(credentials_path)
+    else:
+        raise Exception("No credentials available")
 RAW_DATA_FOLDER= os.getenv('RAW_DATA_FOLDER')
 def chunk_data():
     # Load all JSON files from a directory
diff --git a/services/backend/src/dataflow/rag_model.py b/services/backend/src/dataflow/rag_model.py
index fe05420..e6a2fb9 100644
--- a/services/backend/src/dataflow/rag_model.py
+++ b/services/backend/src/dataflow/rag_model.py
@@ -25,40 +25,37 @@
 FAISS_INDEX_FOLDER= os.getenv('FAISS_INDEX_FOLDER')
 mlflow.set_tracking_uri(MLFLOW_TRACKING_URI)  # Remote MLflow Server
 # Where you currently have this line:
-mlflow.set_experiment("rag_experiment")
-def get_or_create_experiment(experiment_name):
 
-    # Check if experiment exists
+def get_or_create_experiment(experiment_name):
     try:
         experiment = mlflow.get_experiment_by_name(experiment_name)
-        
         if experiment is not None:
-            # Check if experiment is active (not deleted)
             if experiment.lifecycle_stage == "active":
-                print(f"Found active experiment '{experiment_name}' with ID: {experiment.experiment_id}")
+                print(f"✅ Found experiment: {experiment.experiment_id}")
                 return experiment.experiment_id
             else:
-                # Experiment exists but is deleted, create a new one with timestamp
-                new_name = f"{experiment_name}_{int(time.time())}"
-                experiment_id = mlflow.create_experiment(new_name)
-                print(f"Original experiment was deleted. Created new experiment '{new_name}' with ID: {experiment_id}")
-                return experiment_id
-        else:
-            # Create new experiment
-            experiment_id = mlflow.create_experiment(experiment_name)
-            print(f"Created new experiment '{experiment_name}' with ID: {experiment_id}")
-            return experiment_id
-    except Exception as e:
-        print(f"Error getting or creating experiment: {e}")
-        # Fallback - create a new experiment with timestamp
-        new_name = f"{experiment_name}_{int(time.time())}"
-        experiment_id = mlflow.create_experiment(new_name)
-        print(f"Created fallback experiment '{new_name}' with ID: {experiment_id}")
+                print(f"⚠️ Experiment exists but is deleted. Recreating...")
+        # Create a new experiment (either not found or was deleted)
+        experiment_id = mlflow.create_experiment(experiment_name)
+        print(f"🆕 Created experiment '{experiment_name}' with ID: {experiment_id}")
         return experiment_id
+    except Exception as e:
+        print(f"🚨 Exception during experiment creation: {e}")
+        return None
 
 # Replace it with:
-experiment_id = get_or_create_experiment("rag_experiment")
-mlflow.set_experiment_tag("description", "RAG pipeline with Mistral AI model")
+
+def ensure_experiment(name):
+    try:
+        mlflow.set_experiment(name)
+    except Exception as e:
+        mlflow.create_experiment(name)
+        mlflow.set_experiment(name)
+
+ensure_experiment("rag_experiment")
+# mlflow.set_experiment("rag_experiment")
+mlflow.set_tag("description", "RAG pipeline with Mistral AI model")
+# experiment_id = get_or_create_experiment("rag_experiment")
 if not os.environ.get("MISTRAL_API_KEY"):
   os.environ["MISTRAL_API_KEY"] = getpass.getpass("Enter API key for Mistral AI: ")
 
@@ -71,13 +68,21 @@ def get_llm():
 def get_prompt():
 # Define prompt for question-answering
     # Your prompt template
-    template = """Use the following pieces of context to answer the question at the end.
-If you don't know the answer, just say that you don't know, don't try to make up an answer.
-Use three sentences maximum and keep the answer as concise as possible.
-Always say "thanks for asking!" at the end of the answer.
+    template = """You are an expert assistant helping to answer questions based only on the given context.
+
+Instructions:
+- Use the context and search in  below to answer.
+- you can search https://www.khoury.northeastern.edu/ for answering better if not found any in context
+- If context and the website provided  does not contain the answer, say: "I don't know based on the available information."
+- Answer in 2-3 sentences, clearly and factually.
+- End your response with: "Thanks for asking!"
+
+Context:
 {context}
+
 Question: {question}
-Helpful Answer:"""
+
+Answer:"""
     custom_rag_prompt = PromptTemplate.from_template(template)
     return custom_rag_prompt
 
@@ -113,9 +118,9 @@ def load_embeddings():
 vector_store = FAISS.load_local(FAISS_INDEX_FOLDER, embeddings, allow_dangerous_deserialization=True)
 # Define application steps
 def retrieve(state: State):
-    with mlflow.start_run(nested=True, run_name="retrieval",experiment_id=experiment_id):
+    with mlflow.start_run(nested=True, run_name="retrieval"):
         start_time = time.time()
-        retrieved_docs = vector_store.similarity_search(state["question"])
+        retrieved_docs = vector_store.similarity_search(state["question"],k=10)
         retrieval_time = time.time() - start_time
     
         # Extract only metadata
@@ -134,7 +139,7 @@ def retrieve(state: State):
 # Initialize prompt once and store in a global variable
 prompt = get_prompt()
 def generate(state: State):
-    with mlflow.start_run(nested=True, run_name="generation",experiment_id=experiment_id):
+    with mlflow.start_run(nested=True, run_name="generation"):
         start_time = time.time()
         docs_content = "\n\n".join(doc.page_content for doc in state["context"])
         token_count = len(docs_content.split()) 
@@ -161,7 +166,8 @@ def generate(state: State):
 def generateResponse(query):
 # Compile application and test
     try:
-         with mlflow.start_run(run_name="RAG_Pipeline",experiment_id=experiment_id):
+        ensure_experiment("rag_experiment")
+        with mlflow.start_run(run_name="RAG_Pipeline"):
             mlflow.log_param("query", query)
             graph_builder = StateGraph(State).add_sequence([retrieve, generate])
             graph_builder.add_edge(START, "retrieve")
@@ -185,7 +191,12 @@ async def checkModel_fairness():
 if __name__ == "__main__":
 
     query=input("generate query")
+    ensure_experiment("rag_experiment")
     response=generateResponse(query)
+    print("MLflow URI:", mlflow.get_tracking_uri())
+    print("Using experiment ID:", experiment_id)
+    print("Experiments available:", mlflow.search_experiments())
+
     print(response)
     #uncomment and enter prompts for model fairness and there is a limitation on api key
     # asyncio.run(checkModel_fairness())
\ No newline at end of file
diff --git a/services/backend/src/dataflow/scraper.py b/services/backend/src/dataflow/scraper.py
index 69cf26b..0003587 100644
--- a/services/backend/src/dataflow/scraper.py
+++ b/services/backend/src/dataflow/scraper.py
@@ -6,14 +6,30 @@
 import re
 from urllib.parse import urljoin, urlparse
 from dotenv import load_dotenv
-
+import hashlib
 from store_data import upload_many_blobs_with_transfer_manager
 load_dotenv(override=True)
 # Configuration
-BASE_URL = os.getenv('BASE_URL')
+URLS_LIST=list(os.getenv('URLS_LIST','').split(","))
+
+# BASE_URL ="" #URLS_LIST[0]#os.getenv('BASE_URL')
 MAX_DEPTH = int(os.getenv('MAX_DEPTH'))             # Maximum recursion depth (base URL is depth 0)
 CONCURRENT_REQUESTS = int(os.getenv('CONCURRENT_REQUESTS'))  # Maximum number of concurrent requests
-GOOGLE_APPLICATION_CREDENTIALS =os.getenv('GOOGLE_APPLICATION_CREDENTIALS ')
+
+from google.auth import default
+from google.oauth2 import service_account
+
+# Try to get credentials - works in both Docker and Cloud Run
+try:
+    # First try Application Default Credentials (works in Cloud Run)
+    credentials, project = default()
+except Exception:
+    # Fall back to explicit credentials file (for Docker)
+    credentials_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
+    if credentials_path:
+        credentials = service_account.Credentials.from_service_account_file(credentials_path)
+    else:
+        raise Exception("No credentials available")
 # Create folder for JSON data
 DATA_FOLDER = "scraped_data"
 if not os.path.exists(DATA_FOLDER):
@@ -24,7 +40,8 @@ def safe_filename(url):
     parsed = urlparse(url)
     path = parsed.path.strip('/') or 'index'
     filename = re.sub(r'[^A-Za-z0-9_\-]', '_', path) + ".json"
-    return os.path.join(DATA_FOLDER, filename)
+    url_hash = hashlib.md5(url.encode()).hexdigest()[:8]
+    return os.path.join(DATA_FOLDER, f"{filename}_{url_hash}.json")
 
 async def fetch(session, url, semaphore):
     """Fetch the content of the URL asynchronously."""
@@ -39,7 +56,7 @@ async def fetch(session, url, semaphore):
         print(f"Error fetching {url}: {e}")
         return None
 
-async def async_scrape(url, depth=0, session=None, semaphore=None):
+async def async_scrape(url,BASE_URL, depth=0, session=None, semaphore=None):
     """Recursively scrape pages asynchronously and store in JSON format."""
     if depth > MAX_DEPTH:
         return
@@ -81,26 +98,33 @@ async def async_scrape(url, depth=0, session=None, semaphore=None):
         next_url = urljoin(url, link['href'])
         if urlparse(next_url).netloc == urlparse(BASE_URL).netloc:
             next_url = next_url.split('#')[0]  # Remove fragments
-            tasks.append(async_scrape(next_url, depth + 1, session, semaphore))
+            tasks.append(async_scrape(next_url,BASE_URL, depth + 1, session, semaphore))
 
     if tasks:
         await asyncio.gather(*tasks)
     
 
-async def scrape_and_load():
+async def scrape_and_load(CURRENT_URl):
     """Main function to initiate scraping."""
     semaphore = asyncio.Semaphore(CONCURRENT_REQUESTS)
     
     async with aiohttp.ClientSession() as session:
-        await async_scrape(BASE_URL, depth=0, session=session, semaphore=semaphore)
+        await async_scrape(CURRENT_URl,BASE_URL=CURRENT_URl, depth=0, session=session, semaphore=semaphore)
     
 
 def scrape_and_load_task():
-    asyncio.run(scrape_and_load())
+    for url in URLS_LIST:
+        BASE_URL=url
+        asyncio.run(scrape_and_load(BASE_URL))
+        print("*"*15)
+        print(f"scraping {url} done")
+        print("*"*15)
+
     upload_many_blobs_with_transfer_manager()
     return
 
 
 if __name__ == '__main__':
-    asyncio.run(scrape_and_load())
-    upload_many_blobs_with_transfer_manager()
\ No newline at end of file
+    scrape_and_load_task()
+    # asyncio.run(scrape_and_load())
+    # upload_many_blobs_with_transfer_manager()
\ No newline at end of file
diff --git a/services/backend/src/dataflow/store_data.py b/services/backend/src/dataflow/store_data.py
index d142be3..e89ad85 100644
--- a/services/backend/src/dataflow/store_data.py
+++ b/services/backend/src/dataflow/store_data.py
@@ -5,8 +5,20 @@
 BUCKET_NAME= os.getenv('BUCKET_NAME')
 RAW_DATA_FOLDER= os.getenv('RAW_DATA_FOLDER')
 FAISS_INDEX_FOLDER= os.getenv('FAISS_INDEX_FOLDER')
-GOOGLE_APPLICATION_CREDENTIALS=os.getenv('GOOGLE_APPLICATION_CREDENTIALS')
+from google.auth import default
+from google.oauth2 import service_account
 
+# Try to get credentials - works in both Docker and Cloud Run
+try:
+    # First try Application Default Credentials (works in Cloud Run)
+    credentials, project = default()
+except Exception:
+    # Fall back to explicit credentials file (for Docker)
+    credentials_path = os.getenv("GOOGLE_APPLICATION_CREDENTIALS")
+    if credentials_path:
+        credentials = service_account.Credentials.from_service_account_file(credentials_path)
+    else:
+        raise Exception("No credentials available")
 def get_blob_from_bucket():
     storage_client = Client()
     bucket = storage_client.bucket(BUCKET_NAME)
@@ -56,7 +68,7 @@ def upload_many_blobs_with_transfer_manager(
 
     storage_client = Client()
     bucket = storage_client.bucket(BUCKET_NAME)
-    source_directory=os.path.join("..","..","scraped_data")
+    source_directory=os.path.join("scraped_data")
     filenames = [f for f in os.listdir(source_directory) if f.endswith(".json")]
     for filename in filenames:
         file_path = os.path.join(source_directory, filename)
diff --git a/services/frontend/.dockerignore b/services/frontend/.dockerignore
new file mode 100644
index 0000000..4f509e5
--- /dev/null
+++ b/services/frontend/.dockerignore
@@ -0,0 +1 @@
+*.env
\ No newline at end of file
diff --git a/services/frontend/.env b/services/frontend/.env
index 6148832..7a2998f 100644
--- a/services/frontend/.env
+++ b/services/frontend/.env
@@ -1 +1 @@
-API_URL="http://localhost:5002"
\ No newline at end of file
+API_URL="https://backend-service-640053216184.us-east1.run.app/NuBot/"
\ No newline at end of file
diff --git a/services/frontend1/.env b/services/frontend1/.env
index afe9281..77dfb3a 100644
--- a/services/frontend1/.env
+++ b/services/frontend1/.env
@@ -1 +1 @@
-REACT_APP_API_URL=http://localhost:8000/api/chat
\ No newline at end of file
+REACT_APP_API_URL=https://backend-service-640053216184.us-east1.run.app/NuBot/
\ No newline at end of file
diff --git a/services/frontend1/package.json b/services/frontend1/package.json
new file mode 100644
index 0000000..6a2a3fc
--- /dev/null
+++ b/services/frontend1/package.json
@@ -0,0 +1,32 @@
+{
+  "name": "nubot-frontend",
+  "version": "1.0.0",
+  "description": "NUBot React Frontend",
+  "main": "index.js",
+  "scripts": {
+    "start": "react-scripts start",
+    "build": "react-scripts build",
+    "test": "react-scripts test",
+    "eject": "react-scripts eject"
+  },
+  "dependencies": {
+    "axios": "^1.6.2",
+    "react": "^18.2.0",
+    "react-dom": "^18.2.0",
+    "react-router-dom": "^6.20.0",
+    "react-scripts": "5.0.1",
+    "uuid": "^9.0.1"
+  },
+  "browserslist": {
+    "production": [
+      ">0.2%",
+      "not dead",
+      "not op_mini all"
+    ],
+    "development": [
+      "last 1 chrome version",
+      "last 1 firefox version",
+      "last 1 safari version"
+    ]
+  }
+}
\ No newline at end of file
diff --git a/services/frontend1/public/index.html b/services/frontend1/public/index.html
new file mode 100644
index 0000000..b21e2b4
--- /dev/null
+++ b/services/frontend1/public/index.html
@@ -0,0 +1,21 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <!-- Replace the default favicon with your SVG icon -->
+    <link rel="icon" href="%PUBLIC_URL%/nu_tab_icon.svg" type="image/svg+xml" />
+    <!-- Keep the .ico as fallback for browsers that don't support SVG favicons -->
+    <link rel="alternate icon" href="%PUBLIC_URL%/favicon.ico" />
+    <meta name="viewport" content="width=device-width, initial-scale=1" />
+    <meta name="theme-color" content="#000000" />
+    <meta
+      name="description"
+      content="NUBot - Your assistant for all things Northeastern!"
+    />
+    <title>NUBot</title>
+  </head>
+  <body>
+    <noscript>You need to enable JavaScript to run this app.</noscript>
+    <div id="root"></div>
+  </body>
+</html>
diff --git a/services/frontend1/src/ChatInterface.js b/services/frontend1/src/ChatInterface.js
index e48f831..e696e38 100644
--- a/services/frontend1/src/ChatInterface.js
+++ b/services/frontend1/src/ChatInterface.js
@@ -4,20 +4,8 @@ import NULogo from './NULogo';
 import BotAvatar from './BotAvatar';
 import UserAvatar from './UserAvatar';
 
-// Sample chat responses for the full chat interface
-const CHAT_RESPONSES = {
-  "hello": "Hi there! I'm NUBot. How can I help you with Northeastern University information today?",
-  "hi": "Hello! I'm NUBot, your Northeastern University assistant. What information can I help you find?",
-  "courses": "Northeastern offers a wide range of courses across various disciplines. Are you looking for courses in a specific department or program?",
-  "faculty": "Northeastern has renowned faculty members across all colleges. Which department or professor are you interested in learning about?",
-  "campus": "Northeastern's main campus is located in Boston, MA. We also have regional campuses in Charlotte, Seattle, San Francisco, Vancouver, Portland ME, and more. Which campus would you like to know more about?",
-  "about": "I'm NUBot, an AI assistant designed to help you navigate Northeastern University information more easily. I can answer questions about courses, faculty, campus resources, and more!",
-  "help": "I can help you find information about Northeastern's academic programs, faculty, campus resources, student services, and more. What would you like to know?",
-  "admission": "Northeastern has different application processes for undergraduate, graduate, and professional programs. Would you like information about a specific program's admission requirements?",
-  "events": "Northeastern hosts various events and activities throughout the year. You can check the university calendar or specific department pages for upcoming events. Is there a particular type of event you're interested in?",
-  "registration": "Course registration typically opens several months before the start of each semester. The exact dates depend on your student status and program. Would you like to know more about the registration process?",
-  "housing": "Northeastern offers various on-campus housing options for students, from traditional residence halls to apartment-style accommodations. Off-campus housing resources are also available through the university. Would you like specific information about housing options?",
-};
+// Load API URL from environment variable
+const API_URL = process.env.REACT_APP_API_URL || 'http://localhost:8080/NuBot';
 
 // Typing indicator component that uses the three dots
 const TypingIndicator = () => {
@@ -47,7 +35,7 @@ const ChatInterface = () => {
     }
   }, [messages, isLoading]); // Also scroll when loading state changes
   
-  const handleSendMessage = () => {
+  const handleSendMessage = async () => {
     if (!userInput.trim() || isLoading) return;
     
     // Add user message to chat
@@ -58,22 +46,33 @@ const ChatInterface = () => {
     // Set loading state
     setIsLoading(true);
     
-    // Process bot response with delay to simulate thinking/processing
-    setTimeout(() => {
-      const lowerCaseInput = userMessage.toLowerCase();
-      let botResponse = "I'm here to help with anything related to Northeastern University. What else would you like to know?";
+    try {
+      // Make API call to get response
+      const response = await fetch(`${API_URL}`, {
+        method: 'POST',
+        headers: {
+          'Content-Type': 'application/json',
+        },
+        body: JSON.stringify({ query: userMessage }),
+      });
       
-      // Check for keyword matches
-      for (const [keyword, response] of Object.entries(CHAT_RESPONSES)) {
-        if (lowerCaseInput.includes(keyword)) {
-          botResponse = response;
-          break;
-        }
+      if (!response.ok) {
+        throw new Error(`API request failed with status: ${response.status}`);
       }
       
-      setMessages(prev => [...prev, { sender: 'bot', message: botResponse }]);
+      const data = await response.json();
+      // The Flask backend returns the response directly, not wrapped in a field
+      setMessages(prev => [...prev, { sender: 'bot', message: data }]);
+    } catch (error) {
+      console.error('Error fetching response:', error);
+      // Add error message to chat
+      setMessages(prev => [...prev, {
+        sender: 'bot',
+        message: "I'm sorry, I'm having trouble connecting to my knowledge base. Please try again later."
+      }]);
+    } finally {
       setIsLoading(false);
-    }, 1500); // Delay to simulate processing - adjust as needed
+    }
   };
   
   const handleKeyPress = (e) => {